A large retrospective cohort study revealed weight loss and glycemic control independently reduced the risk of various clinical outcomes in patients with type 2 diabetes treated with antidiabetic medications, including glucagon-like peptide-1 receptor agonists. The study used real-world electronic health record data from patients at the Cleveland Clinic, highlighting practical implications beyond controlled clinical trial settings.
The study, published in Diabetes, Obesity and Metabolism, was conducted by researchers from the Cleveland Clinic and Novo Nordisk. They analyzed electronic health record (EHR) data from two distinct patient groups: 8,876 newly diagnosed type 2 diabetes (T2D) patients treated with one of six different antidiabetic drug classes (Cohort A), and 4,161 T2D patients who began treatment with GLP-1 receptor agonists (Cohort B). To improve accuracy, the study applied strict exclusion criteria, removing patients with factors that could influence weight loss, such as organ transplants or chemotherapy.
Key findings:
In Cohort A:
- Each 1% body mass index (BMI) reduction was associated with:
- 3% reduced risk of heart failure (P = .017)
- 1% reduced risk of hypertension (P = .006)
- 4% reduced risk of insulin initiation (P = .001)
- Each 1% (11 mmol/mol) HbA1c reduction was associated with:
- 4% reduced risk of hypertension (P = .041)
- 29% reduced risk of insulin initiation (P = .001)
In Cohort B:
- Each 1% BMI reduction was associated with:
- 4% reduced risk of cardiovascular disease (P = .008)
- 3% reduced risk of insulin initiation (P = .002)
- Each 1% (11 mmol/mol) HbA1c reduction was associated with:
- 4% reduced risk of chronic kidney disease (P = .014)
- 16% reduced risk of insulin initiation (P < .0001)
The study explored the nonlinear relationships between BMI, HbA1c, and clinical outcomes using restricted cubic spline Cox proportional hazard models, adjusting for confounders such as age, sex, race, baseline BMI, baseline HbA1c, and comorbidities.
Additional Results:
In Cohort A:
- Durable weight loss (>5% over 1 year) was associated with:
- 24% reduced risk of nonalcoholic steatohepatitis (P = .036)
- 55% reduced risk of musculoskeletal pain (P = .003)
- Lower BMI variability was associated with decreased risk of gastroesophageal reflux disease (GERD) (P = .001).
- Lower HbA1c variability was associated with decreased risk of dyslipidemia (P = .037) and GERD (P = .011).
In Cohort B:
- Durable weight loss was associated with:
- 72% reduced risk of osteoarthritis (P = .001)
- 72% reduced risk of chronic kidney disease (P = .016)
- 39% reduced risk of essential hypertension (P = .005)
- Durable HbA1c control was associated with:
- 36% reduced risk of essential hypertension (P = .021).
- Higher BMI variability was associated with increased risk of cardiovascular disease (P < .0001) and osteoarthritis (P = .014).
- Higher HbA1c variability was associated with increased risk of cardiovascular disease (P = .003), chronic kidney disease (P = .032), and essential hypertension (P < .0001).
Weight Change Statistics:
- Cohort A: Mean BMI change: -2.008 ± 5.18 kg/m²; median BMI change: -1.447 kg/m².
- Cohort B: Mean BMI change: -1.967 ± 4.8 kg/m²; median BMI change: -1.29 kg/m².
- In both cohorts, around 20% of patients experienced an average BMI increase of over 2.6 kg/m².
HbA1c Change Statistics:
- Cohort A: Mean HbA1c change: -0.389% ± 1.49% (-4 mmol/mol).
- Cohort B: Mean HbA1c change: -0.29% ± 1.26% (-3 mmol/mol).
Durability of Weight Loss and Glycemic Control:
- In Cohort A, 51.21% had durable weight loss, and 46.43% had durable HbA1c control.
- In Cohort B, 62.00% had durable weight loss, and 44.87% had durable HbA1c control.
Time to Insulin Initiation:
- Cohort A: Mean: 48.5 months; median: 37.9 months.
- Cohort B: Mean: 43.9 months; median: 33.1 months.
The study also found significant negative associations between BMI and high-density lipoprotein (HDL) levels, independent of HbA1c (P = .00012 in Cohort A; P < .0001 in Cohort B), and between HbA1c change and HDL, independent of BMI (P < .0001 in both cohorts). Nonlinear associations between BMI changes and cardiovascular disease, GERD, and chronic kidney disease were observed, as well as interactions between BMI and HbA1c for osteoarthritis (P = .02 in Cohort B).
Limitations:
The study, while valuable in assessing real-world outcomes, had limitations including the use of data from a single hospital system, potential generalizability issues, and a relatively short median follow-up period of 5 years. Additionally, physical activity and dietary data were not included, which could have impacted the weight loss outcomes.
Overall, the study emphasizes the importance of treating T2D patients for both glycemic control and weight management to improve long-term clinical outcomes.
A full conflict of interest statement is available in the study.